Deletions of MGF110-9L and MGF360-9L from African swine fever virus are highly attenuated in swine and confer protection against homologous challenge

African swine fever, caused by a large icosahedral DNA virus (African swine fever virus, ASFV), is a highly contagious disease in domestic and feral swine, thus posing a significant economic threat to the global swine industry. Currently, there are no effective vaccines or the available methods to control ASFV infection. Attenuated live viruses with deleted virulence factors are considered to be the most promising vaccine candidates; however, the mechanism by which these attenuated viruses confer protection is unclear. Here, we used the Chinese ASFV CN/GS/2018 as a backbone and used homologous recombination to generate a virus in which MGF110-9L and MGF360-9L, two genes antagonize host innate antiviral immune response, were deleted (ASFV-ΔMGF110/360-9L). This genetically modified virus was highly attenuated in pigs and provided effective protection of pigs against parental ASFV challenge. Importantly, we found ASFV-ΔMGF110/360-9L infection induced higher expression of Toll-like receptor 2 (TLR2) mRNA compared with parental ASFV as determined by RNA-Seq and RT-PCR analysis. Further immunoblotting results showed that parental ASFV and ASFV-ΔMGF110/360-9L infection inhibited Pam3CSK4-triggered activating phosphorylation of proinflammatory transcription factor NF-κB subunit p65 and phosphorylation of NF-κB inhibitor IκBα levels, although NF-κB activation was higher in ASFV-ΔMGF110/360-9L-infected cells compared with parental ASFV-infected cells. Additionally, we show overexpression of TLR2 inhibited ASFV replication and the expression of ASFV p72 protein, whereas knockdown of TLR2 had the opposite effect. Our findings suggest that the attenuated virulence of ASFV-ΔMGF110/360-9L might be mediated by increased NF-κB and TLR2 signaling.

African swine fever, caused by a large icosahedral DNA virus (African swine fever virus, ASFV), is a highly contagious disease in domestic and feral swine, thus posing a significant economic threat to the global swine industry. Currently, there are no effective vaccines or the available methods to control ASFV infection. Attenuated live viruses with deleted virulence factors are considered to be the most promising vaccine candidates; however, the mechanism by which these attenuated viruses confer protection is unclear. Here, we used the Chinese ASFV CN/GS/2018 as a backbone and used homologous recombination to generate a virus in which MGF110-9L and MGF360-9L, two genes antagonize host innate antiviral immune response, were deleted (ASFV-ΔMGF110/360-9L). This genetically modified virus was highly attenuated in pigs and provided effective protection of pigs against parental ASFV challenge. Importantly, we found ASFV-ΔMGF110/360-9L infection induced higher expression of Toll-like receptor 2 (TLR2) mRNA compared with parental ASFV as determined by RNA-Seq and RT-PCR analysis. Further immunoblotting results showed that parental ASFV and ASFV-ΔMGF110/360-9L infection inhibited Pam3CSK4-triggered activating phosphorylation of proinflammatory transcription factor NF-κB subunit p65 and phosphorylation of NF-κB inhibitor IκBα levels, although NF-κB activation was higher in ASFV-ΔMGF110/360-9L-infected cells compared with parental ASFV-infected cells. Additionally, we show overexpression of TLR2 inhibited ASFV replication and the expression of ASFV p72 protein, whereas knockdown of TLR2 had the opposite effect. Our findings suggest that the attenuated virulence of ASFV-ΔMGF110/360-9L might be mediated by increased NF-κB and TLR2 signaling.
African swine fever virus (ASFV) is the causative agent of African swine fever (ASF), a contagious and often fatal disease of domestic pigs and wild boar. ASFV has a complicated architecture with a linear, double-stranded DNA genome (170-193 kb) containing 150 to 167 ORFs (1)(2)(3). Currently, ASF is endemic in more than 20 sub-Saharan African countries with serious economic consequences for the swine industry (3,4). The disease spread from Africa to the Caucasus, especially within the country of Georgia in 2007 (5). Afterward, ASF spread to most countries in the eastern part of the European Union in 2014. In August 2018, the ASF epidemic started in China and it has caused huge economic losses to the pig industry (6).
Different vaccine strategies for ASF have been evaluated in the past decades. However, due to the huge ASFV genome and complex mutation mechanism, no clinically effective vaccines have been developed to date (7), including inactivated vaccines, DNA vaccines, subunit vaccines, and adenovirusvectored vaccines (8)(9)(10)(11)(12). Rational development of attenuated strains by genetic manipulation is a valid alternative, and perhaps a safer methodology, compared to the use of naturally attenuated isolates (3,13). Some attenuated ASFV strains, obtained by genetic manipulation, consisting of deletions of single genes or a group of genes, have shown to induce protection against the virulent parental virus (3,(14)(15)(16)(17)(18)(19)(20)(21)(22); however, the mechanism of attenuation and protection is unclear.

Construction of ASFV-ΔMGF110/360-9L virus
ASFV-encoding genes MGF110-9L (23) and MGF360-9L (25) have been identified as the virulence factor of ASFV. To further determine the combined roles of MGF110-9L with MGF360-9L during ASFV infection in cell cultures and the virulence and pathogenicity in swine, a recombinant virus (ASFV-ΔMGF110/360-9L) lacking the MGF110-9L and MGF360-9L genes was constructed from a highly pathogenic genotype II virus (ASFV CN/GS/2018) using homologous recombination. The MGF110-9L and MGF360-9L genes were replaced by a cassette containing the fluorescent gene mCherry or eGFP, respectively, under the ASFV p72 promoter (Fig. 1A). The recombinant virus was obtained after ten successive rounds of limiting dilution purification based on fluorescence activity. The recombinant virus obtained from the last round of purification was amplified in PAMs to obtain a virus stock. Red and green fluorescence were observed in ASFV-ΔMGF110/360-9L-infected PAM cells at 12 h postinfection (Fig. 1B). The absence of MGF110-9L and MGF360-9L genes in the ASFV-ΔMGF110/360-9L mutant virus was confirmed by sequence analysis of both the parental and mutant viruses and confirmed by PCR. MGF110-9L or MGF360-9L genes were detected in ASFV-infected PAMs but not in ASFV-ΔMGF110/360-9L-infected PAMs, and the p72 gene was detected in ASFV-infected PAMs and ASFV-ΔMGF110/360-9L-infected PAMs (Figs. 1C and S1). Collectively, these results indicate that the ASFV-ΔMGF110/360-9L mutant virus was successfully constructed.

TLR2 inhibited ASFV replication in PAMs
To investigate the effect of TLR2 on ASFV replication, Ma-104 cells were transfected with HA-TLR2 expression plasmid and then infected with ASFV. As shown in Figure 7A and Fig. S3, ectopically expressed TLR2 reduced ASFV titer in Ma-104 cells. Consistently, expression of ASFV p72 also decreased in TLR2-overexpressing Ma-104 cells in comparison with that of control cells (Fig. 7B). Moreover, TLR2 expression was silenced in PAMs using a small TLR2 interfering RNA (RNAi), and the results showed that TLR2 expression was markedly reduced in ASFV-infected PAM cells (Fig. 7C). Consequently, the ASFV titer in TLR2-knockdown cells was higher than that in control cells (Figs. 7D and Fig. S4) at early ASFV infection, and ASFV p72 protein expression was higher in TLR2knockdown PAM cells than that in control cells (Fig. 7E). These data suggest that TLR2 inhibited ASFV replication.

Discussion
There is a lack of information on the function of many ASFV genes. Deletions of single genes or a group of genes to obtain attenuated strains have been shown to induce protection against the virulent parental virus (14,15,19,20,22,26,27). In this study, MGF110-9L and MGF360-9L genes were identified as genetic determinants of virulence as their deletions of these genes highly attenuates the pathogenicity of ASFV CN/GS/2018 up to an HAD 50 dose of 10 4 in pigs.
Virulence may be completely abolished in the highly virulent genotype II ASFV by genetic manipulation. For instance, deletion of the 9 GL gene (particularly potentiated by the additional deletion of the UK gene), deletion of six genes from the MGF360 and MGF530 regions, deletion of the I177L gene, or deletion of seven genes encoding MGF505-1R, MGF505-2R, MGF505-3R, MGF360-12L, MGF360-13L, MGF360-14L, and E402R caused the attenuation and complete protection of pigs against lethal ASFV challenge (3,(19)(20)(21). Our results showed that pigs infected with 10 4 HAD 50 of ASFV-ΔMGF110/360-9L  were effectively protected when challenged at 17 dpi with the virulent parental strain with no disease-associated signs, indicating the safety of ASFV-ΔMGF110/360-9L as a potential vaccine candidate. Furthermore, it appears that replication of the challenge virus in the ASFV-ΔMGF110/360-9L-infected pigs was restricted.
The host mechanisms mediating protection against ASFV infection remain unclear. ASFV E66L inhibits host translation by the PKR/eIF2α pathway (28). DP96R of ASFV China 2018/1 negatively regulates type I IFN expression and NF-κB signaling by inhibiting both TBK1 and IKKβ, which play an important role in virus immune evasion (29). ASFV MGF360-12L inhibits type I IFN production by blocking the interaction of importinα and the NF-κB signaling pathway (30). Besides, ASFV encoding E120R protein has been confirmed to suppress type I IFN production through interacting with IRF3 (31); however, there is no evaluation of virulence and protection against parental ASFV challenge in swine by deletion of those genes. Deletion of ASFV MGF505-7R is highly attenuated in pigs at low dose by upregulating STING expression (32). I267L deficiency attenuates the virulence and pathogenesis of ASFV in pigs by impairing innate immune responses mediated by the RNA Pol-III-RIG-I axis (33); however, there is no evaluation of protection against parental ASFV challenge in swine. Analysis of different live attenuated vaccine candidates shows that there is a close association between the presence of circulating virusspecific antibodies and protection (3). However, a limited number of samples were tested; therefore, further study is required to determine whether this is the case for other vaccines.
In this study, the deletions of MGF110-9L and MGF360-9L are highly attenuated in pigs at high dose and effectively protect the pigs against lethal ASFV challenge. Generally, the protection of pigs caused by ASFV-deleted strains is multifactorial. Our previous study illustrated that MGF110-9L gene is transcribed at an early stage in the virus replication cycle (23) and is the inhibitor of type I IFN production (unpublished data). Besides, the upregulation of chemokines [e.g., C-X-C Motif Chemokine Ligand 10 (CXCL10), C-C Motif Chemokine Ligand 2 (CCL2) and C-C Motif Chemokine Ligand 11 (CCL11)] and pro-inflammatory cytokines (e.g., IL1β, IL1α, and IL13) can be observed in the ASFV-ΔMGF110-9L-infected PAMs (unpublished data), indicating the role of MGF110-9L on host innate immune response and inflammatory response. Furthermore, the difference can be found between parental ASFV and ASFV-ΔMGF110-9L infection in the regulation of MHC class II protein binding, Toll signaling pathway, apoptosis, arginine and proline metabolism and purine metabolism (unpublished data). These results implied that MGF110-9L may be involved in the regulation of multiple host functions. Our previous study also screened several host factors [Proteasome subunit alpha type (PSMA3), 26S protease regulatory subunit 4 (PSMC1), autophagy and beclin 1 regulator 1 (AMBRA1), and DEAD-box helicase 20 (DDX20)] interacted with MGF360-9L by constructing the interaction network of MGF360-9L with host proteins (34). Of these, PSMA3 and PSMC1 involved in proteasome catabolism have been verified to promote ASFV replication (34). These results indicated that MGF360-9L regulates host-cell biological processes through the proteasome pathway.
To explore the potential cause of protective phenomenon mediated by combinational deletions of MGF110-9L and MGF360-9L, functional analysis was performed, and we observed DEGs between ASFV and ASFV-ΔMGF110/360-9Linfected PAMs were mainly enriched in immune response and inflammatory response. MGF360-9L has been reported to inhibit host innate immunity by degrading STAT1 and STAT2 (24), and MGF110-9L is also an inhibitor of type I IFN production (unpublished data). Therefore, it is obvious that the enhanced innate immune response caused by ASFV-ΔMGF110/360-9L is one of a factor in the protection of pigs. We next investigate the role of inflammatory response in protection of pigs. We first observed that ASFV-ΔMGF110/ 360-9L induced more Tlr2 mRNA than ASFV by RNA-Seq and qPCR assay. TLR2, one of the pattern recognition receptors expressed on the surface of immune cells, is known as a sensor of bacterial lipoproteins. Besides, TLR2 has also been reported to sense molecular patterns of viruses such as hepatitis C virus, vaccinia virus, and respiratory syncytial virus (35)(36)(37)(38)(39)(40). Engagement of the TLR2 axis leads to activation of NF-κB pathway, increased gene expression, and release of inflammation-driving mediators such as inter alia IL1β and TNFα. In this study, ASFV-ΔMGF110/360-9L enhanced Pam3CSK4-triggered TLR2 signaling compared to that of parental ASFV, indicating ASFV-ΔMGF110/360-9L induced stronger inflammatory response than parental ASFV. Finally, overexpression of TLR2 inhibited ASFV replication. Conversely, knockdown of TLR2 by RNAi in PAMs increased ASFV replication. These results demonstrated that inflammatory responses mediated by TLR2 signaling are also another factor in protection of pigs.
In summary, the Chinese ASFV CN/GS/2018 was used as a backbone to generate a virus bearing MGF110-9L and MGF360-9L gene deletions, which is highly attenuated in pigs, has a low risk of converting to a virulent strain, and could induce effective protection in pigs against lethal ASFV CN/GS/ 2018 challenge. ASFV-ΔMGF110/360-9L increased host Tlr2 mRNA levels, and TLR2 inhibited ASFV replication. Our results confirmed that the deletions of MGF110-9L and MGF360-9L provide another target for the production of rationally attenuated ASFV vaccine candidates and that the attenuated virulence of ASFV-ΔMGF110/360-9L might be mediated by weakening the inhibitory effect of ASFV-ΔMGF110/360-9L on TLR2 signaling.

Ethics statements
This study was carried out in strict accordance with the recommendations provided in the Guide for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of the People's Republic of China. All protocols were approved by the Committee on the Ethics of Animal Experiments of the Lanzhou Veterinary Research Institute of the Chinese Academy of Agricultural Sciences and the Animal Ethics Committee of Lanzhou Province, China.

Biosafety statement and facility
All experiments involved in ASFV were conducted under the enhanced biosafety level 3 (P3) facilities in the Lanzhou Veterinary Research Institute of the Chinese Academy of Agricultural Sciences and approved by the Ministry of Agriculture and Rural Affairs and the China National Accreditation Service for Conformity Assessment.

Cell culture, viruses, and pigs
PAMs were isolated from 30 to 40 day-old piglets as previously described (41) and maintained in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (BI) at 37 C with 5% CO 2 . ASFV CN/GS/2018 was isolated from field samples in China as described previously and stored in our laboratory (32). Ma-104 cells and iPAM cells (ATCC) were cultured in Dulbecco's Modified Eagle Medium.
Here, 50-day-old healthy Large White-Duroc crossbred pigs were obtained from local farms, with each pig being antigenically and serologically negative for ASFV, porcine respiratory and reproductive syndrome, classical swine fever, and pseudorabies virus.

Construction of recombinant ASFV-ΔMGF110/360-9L
Recombinant ASFV-ΔMGF110/360-9L was produced by homologous recombination between ASFV-ΔMGF360-9L (24) and recombinant transfer vector. Recombinant transfer vector containing the flanking genomic regions of the MGF-110-9L gene mapping to the left (1000 bp) and right (1000 bp) of the gene and a reporter gene cassette containing the mCherry gene with the ASFV p72 late gene promoter was used (Fig. 1). Recombinant transfer vectors were generated by fusing PCR and the Gibson Assembly technique (Invitrogen Life Sciences). PAMs were transfected with the constructed transfer vector using TransITLT1 transfection reagent (Mirus Bio) and then infected with ASFV-ΔMGF360-9L at 24 h post-transfection as described previously (21). The resulting virus ASFV-ΔMGF110/360-9L was purified by successively selecting fluorescent plaques combined with limited dilution on monolayers of PAMs. Virus from the last round of purification was propagated in PAMs to obtain a virus stock. Purity of ASFV-ΔMGF110/360-9L in the virus stock was evaluated by PCR and sequencing utilizing primers MGF110-9L-F/R and MGF360-9L-F/R (Table 1).

Multistep virus growth curves
Growth curves comparing growth kinetics among the parental ASFV CN/GS/2018, recombinant ASFV-ΔMGF360-9L, ASFV-ΔMGF110-9L, and ASFV-ΔMGF110/360-9L were conducted in PAMs as described previously (17,42). Briefly, PAMs were seeded in 48-well plates and infected at an MOI of 0.1. One hour later, the inoculum was removed and the cells were rinsed three times with PBS. PAMs were then rinsed with RPMI 1640 supplemented with 2% fetal bovine serum and incubated for 2, 24, 48, 72, and 96 h at 37 C with 5% CO 2 . At the appropriate times postinfection, cells were collected and frozen at −80 C, and the thawed lysates were used to determine titers by HAD 50 /0.1 ml (42) in PAMs. In this displayed experiment, three replicates were included per experimental condition.
qPCR assay ASFV genomic DNA was extracted from cell supernatants, tissue homogenate, or EDTA-treated whole peripheral blood using GenElute Mammalian Genomic DNA Miniprep Kits (Sigma-Aldrich). qPCR assay was carried out to amplify the conserved p72 gene segment on a QuantStudio 5 real-time PCR system (Applied Biosystems) according to the OIErecommended procedure. A TaqMan probe (5 0 -[6-carboxyfluorescein(FAM)]-ccacgggaggaataccaacccagtg-3 0 -[6-carboxy tetramethyl-rhodamine] from Applied Biosystems) was designed from an alignment of 54 available ASFV sequences for the 3 0 -end of p72. The quantity of ASFV genome was calculated by the standard curve and presented as genome copies per microliter.

Real-time PCR (RT-PCR)
Total RNA was isolated from cells using TRIzol reagent (TaKaRa) and reverse transcribed into the cDNA utilizing PrimeScript RT reagent kit (TaKaRa). RT-PCR analysis was conducted to measure mRNA levels of tested genes. Data shown are the relative abundance of the indicated mRNAs normalized to that of GAPDH. Relative mRNA expression levels were assessed and calculated utilizing 2 ΔΔCT method. Gene-specific primer sequences were listed in Table 1.

Immunoblotting assay
For endogenous immunoblotting assay, PAMs were infected with parental ASFV, recombinant ASFV-ΔMGF110-9L, ASFV-ΔMGF360-9L, or ASFV-ΔMGF110/360-9L at an MOI of 0.1 for 6, 12, and 18 h and then treated with Pam3CSK4 (100 ng/ml) for an additional 4 h. For exogenous assay, iPAM cells were transfected with Flag-MGF360-9L and HA-TLR2 plasmids for 20 h and then treated with Pam3CSK4 (100 ng/ ml) for an additional 4 h. Cells were lysed in lysis buffer, and then the extracted proteins were separated by SDS-PAGE gels.

Animal experiments
To evaluate the virulence of ASFV-ΔMGF110/360-9L in pigs, groups of 50-day-old pigs (n = 6) were i.m. inoculated with 10 4 HAD 50 of either parental ASFV or ASFV-ΔMGF110/ 360-9L. Presence of clinical signs including temperature and mortality were recorded daily for 17 days. To further evaluate the protective efficacy of ASFV-ΔMGF110/360-9L against parental virulent ASFV challenge, pigs inoculated with ASFV-ΔMGF110/360-9L were i.m. challenged 17 days later with 10 2 HAD 50 of parental lethal ASFV. Pigs were monitored daily for 17 days p.c. for clinical signs and mortality. Blood and tissues including the heart, liver, spleen, lung, kidney, and thymus from the dead pigs or surviving pigs, which were euthanized at the end of the observation period, were collected for ruling out the presence of ASFV using qPCR as recommended by OIE (43).

Transcriptome sequencing (RNA-seq) and analysis
PAMs were infected with ASFV or ASFV-ΔMGF110/360-9L for 12 or 24 h, and then total RNA was extracted and used for the reference transcriptome analysis by Oebiotech. Libraries were constructed using a VAHTSTM mRNA-seq V3 Library Prep Kit for Illumina (Vazyme Inc) according to the manufacturer's instructions. In brief, the mRNA separation/fragmentation reaction was performed, and the resulting 200 to 300 bp fragments were inserted using the hold fragmentation program at 4 C. First strand cDNA was synthesized using random primers and reverse transcriptase, and a specific library was established when the second strand was synthesized. The library was amplified, and total DNA was screened to isolate 320 to 420 bp strands. High-throughput single-end 75 bp sequencing was performed using the NextSeq 500 Sequencing System (Illumina). Quantitative mapping was performed using the FANSe2 algorithm, and DESeq software (1.16.1) was used to screen out different genes according to the thresholds of ≥2-fold change in expression and a p value < 0.05. Finally, the Kyoto Encyclopedia of Genes and Genomes pathway and GO enrichment analyses were performed for the DEGs using the DAVID database.
ELISA assay for detection of anti-ASFV antibodies ELISA microtiter plates were coated with 800 ng per well of p30 protein in a 0.05 M carbonate buffer solution (pH 9.6) and incubated overnight at 4 C. Antigen-coated plates were washed three times with PBS containing 0.5% v/v Tween-20 (PBST) and then blocked with 200 μl of blocking buffer (2.5% w/v nonfat dry milk in PBST) overnight at 4 C. After three washes with PBST, 100 μl of test samples, positive serum samples, and negative serum samples diluted in blocking buffer at 1:2 were separately added to each well in duplicate. The plates were then incubated for 30 min at 37 C followed by three washes with PBST, before adding 100 μl/well of p30 mAb-HRP (1:25,000) and incubating at 37 C for an additional 30 min. Following a final triplicate wash with PBST, 100 μl/ well of 3,3 0 ,5,5 0 -tetramethylbenzidine substrate, prepared by mixing solution A (205 mM potassium citrate, pH 4.0) and solution B (41 mM 3,3 0 ,5,5 0 -tetramethylbenzidine) in a 39:1 v/v ratio, was added to each well, and the plates were incubated in the dark for 15 min at 37 C. As a final step, 2 M H 2 SO 4 (50 μl/ well) was used to stop the colorimetric reaction, and the A 450nm values were read using an automated ELISA plate reader.

Statistical analysis
Results are shown as mean ± SD. An unpaired two-tailed Student's t test was used to analyze the significance of the results (p < 0.05). All experiments were performed independently at least three times, with a representative experiment being shown.

Data availability
All data supporting the findings of this study are available within this paper.
Supporting information-This article contains supporting information.